Datasets:
sswt
/
arxiv_sample_50K

Dataset card Data Studio Files
xet
Community
1
text
stringlengths
62
2.94k
Summary Grounded Conversation Generation ; Many conversation datasets have been constructed in the recent years using crowdsourcing. However, the data collection process can be time consuming and presents many challenges to ensure data quality. Since language generation has improved immensely in recent years with the advancement of pretrained language models, we investigate how such models can be utilized to generate entire conversations, given only a summary of a conversation as the input. We explore three approaches to generate summary grounded conversations, and evaluate the generated conversations using automatic measures and human judgements. We also show that the accuracy of conversation summarization can be improved by augmenting a conversation summarization dataset with generated conversations.
Biomedical DatatoText Generation via FineTuning Transformers ; Datatotext D2T generation in the biomedical domain is a promising yet mostly unexplored field of research. Here, we apply neural models for D2T generation to a realworld dataset consisting of package leaflets of European medicines. We show that finetuned transformers are able to generate realistic, multisentence text from data in the biomedical domain, yet have important limitations. We also release a new dataset BioLeaflets for benchmarking D2T generation models in the biomedical domain.
Freiheitssatz and phase transition for the density model of random groups ; Magnus' Freiheitssatz states that if a group is defined by a presentation with m generators and a single relator containing the last generating letter, then the first m1 letters freely generate a free subgroup. We study an analogue of this theorem in the Gromov density model of random groups, showing a phase transition phenomenon at density dr minfrac12, 1log2m12r1 with 1leq rleq m1 we prove that for a random group with m generators at density d, if d dr then the first r letters freely generate a free subgroup; whereas if d dr then the first r letters generate the whole group.
Explore the Expression Facial Expression Generation using Auxiliary Classifier Generative Adversarial Network ; Facial expressions are a form of nonverbal communication that humans perform seamlessly for meaningful transfer of information. Most of the literature addresses the facial expression recognition aspect however, with the advent of Generative Models, it has become possible to explore the affect space in addition to mere classification of a set of expressions. In this article, we propose a generative model architecture which robustly generates a set of facial expressions for multiple character identities and explores the possibilities of generating complex expressions by combining the simple ones.
nBIIG A Neural BI Insights Generation System for Table Reporting ; We present nBIIG, a neural Business Intelligence BI Insights Generation system. Given a table, our system applies various analyses to create corresponding RDF representations, and then uses a neural model to generate fluent textual insights out of these representations. The generated insights can be used by an analyst, via a humanintheloop paradigm, to enhance the task of creating compelling table reports. The underlying generative neural model is trained over large and carefully distilled data, curated from multiple BI domains. Thus, the system can generate faithful and fluent insights over opendomain tables, making it practical and useful.
Black Box Adversarial Prompting for Foundation Models ; Prompting interfaces allow users to quickly adjust the output of generative models in both vision and language. However, small changes and design choices in the prompt can lead to significant differences in the output. In this work, we develop a blackbox framework for generating adversarial prompts for unstructured image and text generation. These prompts, which can be standalone or prepended to benign prompts, induce specific behaviors into the generative process, such as generating images of a particular object or generating high perplexity text.
DiffPattern Layout Pattern Generation via Discrete Diffusion ; Deep generative models dominate the existing literature in layout pattern generation. However, leaving the guarantee of legality to an inexplicable neural network could be problematic in several applications. In this paper, we propose toolDiffPattern to generate reliable layout patterns. toolDiffPattern introduces a novel diverse topology generation method via a discrete diffusion model with computeefficiently lossless layout pattern representation. Then a whitebox pattern assessment is utilized to generate legal patterns given desired design rules. Our experiments on several benchmark settings show that toolDiffPattern significantly outperforms existing baselines and is capable of synthesizing reliable layout patterns.
Trustworthiness of Children Stories Generated by Large Language Models ; Large Language Models LLMs have shown a tremendous capacity for generating literary text. However, their effectiveness in generating children's stories has yet to be thoroughly examined. In this study, we evaluate the trustworthiness of children's stories generated by LLMs using various measures, and we compare and contrast our results with both old and new children's stories to better assess their significance. Our findings suggest that LLMs still struggle to generate children's stories at the level of quality and nuance found in actual stories
Vulnerabilities in AI Code Generators Exploring Targeted Data Poisoning Attacks ; In this work, we assess the security of AI code generators via data poisoning, i.e., an attack that injects malicious samples into the training data to generate vulnerable code. We poison the training data by injecting increasing amounts of code containing security vulnerabilities and assess the attack's success on different stateoftheart models for code generation. Our analysis shows that AI code generators are vulnerable to even a small amount of data poisoning. Moreover, the attack does not impact the correctness of code generated by pretrained models, making it hard to detect.
SGG Learning to Select, Guide, and Generate for Keyphrase Generation ; Keyphrases, that concisely summarize the highlevel topics discussed in a document, can be categorized into present keyphrase which explicitly appears in the source text, and absent keyphrase which does not match any contiguous subsequence but is highly semantically related to the source. Most existing keyphrase generation approaches synchronously generate present and absent keyphrases without explicitly distinguishing these two categories. In this paper, a SelectGuideGenerate SGG approach is proposed to deal with present and absent keyphrase generation separately with different mechanisms. Specifically, SGG is a hierarchical neural network which consists of a pointingbased selector at low layer concentrated on present keyphrase generation, a selectionguided generator at high layer dedicated to absent keyphrase generation, and a guider in the middle to transfer information from selector to generator. Experimental results on four keyphrase generation benchmarks demonstrate the effectiveness of our model, which significantly outperforms the strong baselines for both present and absent keyphrases generation. Furthermore, we extend SGG to a title generation task which indicates its extensibility in natural language generation tasks.
Generative Adversarial Mapping Networks ; Generative Adversarial Networks GANs have shown impressive performance in generating photorealistic images. They fit generative models by minimizing certain distance measure between the real image distribution and the generated data distribution. Several distance measures have been used, such as JensenShannon divergence, fdivergence, and Wasserstein distance, and choosing an appropriate distance measure is very important for training the generative network. In this paper, we choose to use the maximum mean discrepancy MMD as the distance metric, which has several nice theoretical guarantees. In fact, generative moment matching network GMMN Li, Swersky, and Zemel 2015 is such a generative model which contains only one generator network G trained by directly minimizing MMD between the real and generated distributions. However, it fails to generate meaningful samples on challenging benchmark datasets, such as CIFAR10 and LSUN. To improve on GMMN, we propose to add an extra network F, called mapper. F maps both real data distribution and generated data distribution from the original data space to a feature representation space mathcalR, and it is trained to maximize MMD between the two mapped distributions in mathcalR, while the generator G tries to minimize the MMD. We call the new model generative adversarial mapping networks GAMNs. We demonstrate that the adversarial mapper F can help G to better capture the underlying data distribution. We also show that GAMN significantly outperforms GMMN, and is also superior to or comparable with other stateoftheart GAN based methods on MNIST, CIFAR10 and LSUNBedrooms datasets.
Towards Generating Diverse Audio Captions via Adversarial Training ; Automated audio captioning is a crossmodal translation task for describing the content of audio clips with natural language sentences. This task has attracted increasing attention and substantial progress has been made in recent years. Captions generated by existing models are generally faithful to the content of audio clips, however, these machinegenerated captions are often deterministic e.g., generating a fixed caption for a given audio clip, simple e.g., using common words and simple grammar, and generic e.g., generating the same caption for similar audio clips. When people are asked to describe the content of an audio clip, different people tend to focus on different sound events and describe an audio clip diversely from various aspects using distinct words and grammar. We believe that an audio captioning system should have the ability to generate diverse captions, either for a fixed audio clip, or across similar audio clips. To this end, we propose an adversarial training framework based on a conditional generative adversarial network CGAN to improve diversity of audio captioning systems. A caption generator and two hybrid discriminators compete and are learned jointly, where the caption generator can be any standard encoderdecoder captioning model used to generate captions, and the hybrid discriminators assess the generated captions from different criteria, such as their naturalness and semantics. We conduct experiments on the Clotho dataset. The results show that our proposed model can generate captions with better diversity as compared to stateoftheart methods.
PV3D A 3D Generative Model for Portrait Video Generation ; Recent advances in generative adversarial networks GANs have demonstrated the capabilities of generating stunning photorealistic portrait images. While some prior works have applied such image GANs to unconditional 2D portrait video generation and static 3D portrait synthesis, there are few works successfully extending GANs for generating 3Daware portrait videos. In this work, we propose PV3D, the first generative framework that can synthesize multiview consistent portrait videos. Specifically, our method extends the recent static 3Daware image GAN to the video domain by generalizing the 3D implicit neural representation to model the spatiotemporal space. To introduce motion dynamics to the generation process, we develop a motion generator by stacking multiple motion layers to generate motion features via modulated convolution. To alleviate motion ambiguities caused by camerahuman motions, we propose a simple yet effective camera condition strategy for PV3D, enabling both temporal and multiview consistent video generation. Moreover, PV3D introduces two discriminators for regularizing the spatial and temporal domains to ensure the plausibility of the generated portrait videos. These elaborated designs enable PV3D to generate 3Daware motionplausible portrait videos with highquality appearance and geometry, significantly outperforming prior works. As a result, PV3D is able to support many downstream applications such as animating static portraits and viewconsistent video motion editing. Code and models are released at httpsshowlab.github.iopv3d.
Quantum Gravity with Matter via Group Field Theory ; A generalization of the matrix model idea to quantum gravity in three and higher dimensions is known as group field theory GFT. In this paper we study generalized GFT models that can be used to describe 3D quantum gravity coupled to point particles. The generalization considered is that of replacing the group leading to pure quantum gravity by the twisted product of the group with its dual the socalled Drinfeld double of the group. The Drinfeld double is a quantum group in that it is an algebra that is both noncommutative and noncocommutative, and special care is needed to define group field theory for it. We show how this is done, and study the resulting GFT models. Of special interest is a new topological model that is the PonzanoRegge'' model for the Drinfeld double. However, as we show, this model does not describe point particles. Motivated by the GFT considerations, we consider a more general class of models that are defined using not GFT, but the socalled chain mail techniques. A general model of this class does not produce 3manifold invariants, but has an interpretation in terms of point particle Feynman diagrams.
Efficient generative modeling of protein sequences using simple autoregressive models ; Generative models emerge as promising candidates for novel sequencedata driven approaches to protein design, and for the extraction of structural and functional information about proteins deeply hidden in rapidly growing sequence databases. Here we propose simple autoregressive models as highly accurate but computationally efficient generative sequence models. We show that they perform similarly to existing approaches based on Boltzmann machines or deep generative models, but at a substantially lower computational cost by a factor between 102 and 103. Furthermore, the simple structure of our models has distinctive mathematical advantages, which translate into an improved applicability in sequence generation and evaluation. Within these models, we can easily estimate both the probability of a given sequence, and, using the model's entropy, the size of the functional sequence space related to a specific protein family. In the example of response regulators, we find a huge number of ca. 1068 possible sequences, which nevertheless constitute only the astronomically small fraction 1080 of all aminoacid sequences of the same length. These findings illustrate the potential and the difficulty in exploring sequence space via generative sequence models.
GlueGen Plug and Play Multimodal Encoders for Xtoimage Generation ; Texttoimage T2I models based on diffusion processes have achieved remarkable success in controllable image generation using userprovided captions. However, the tight coupling between the current text encoder and image decoder in T2I models makes it challenging to replace or upgrade. Such changes often require massive finetuning or even training from scratch with the prohibitive expense. To address this problem, we propose GlueGen, which applies a newly proposed GlueNet model to align features from singlemodal or multimodal encoders with the latent space of an existing T2I model. The approach introduces a new training objective that leverages parallel corpora to align the representation spaces of different encoders. Empirical results show that GlueNet can be trained efficiently and enables various capabilities beyond previous stateoftheart models 1 multilingual language models such as XLMRoberta can be aligned with existing T2I models, allowing for the generation of highquality images from captions beyond English; 2 GlueNet can align multimodal encoders such as AudioCLIP with the Stable Diffusion model, enabling soundtoimage generation; 3 it can also upgrade the current text encoder of the latent diffusion model for challenging case generation. By the alignment of various feature representations, the GlueNet allows for flexible and efficient integration of new functionality into existing T2I models and sheds light on Xtoimage X2I generation.
Generative Model for Models Rapid DNN Customization for Diverse Tasks and Resource Constraints ; Unlike cloudbased deep learning models that are often large and uniform, edgedeployed models usually demand customization for domainspecific tasks and resourcelimited environments. Such customization processes can be costly and timeconsuming due to the diversity of edge scenarios and the training load for each scenario. Although various approaches have been proposed for rapid resourceoriented customization and taskoriented customization respectively, achieving both of them at the same time is challenging. Drawing inspiration from the generative AI and the modular composability of neural networks, we introduce NNFactory, an oneforall framework to generate customized lightweight models for diverse edge scenarios. The key idea is to use a generative model to directly produce the customized models, instead of training them. The main components of NNFactory include a modular supernet with pretrained modules that can be conditionally activated to accomplish different tasks and a generative module assembler that manipulate the modules according to task and sparsity requirements. Given an edge scenario, NNFactory can efficiently customize a compact model specialized in the edge task while satisfying the edge resource constraints by searching for the optimal strategy to assemble the modules. Based on experiments on image classification and object detection tasks with different edge devices, NNFactory is able to generate highquality task and resourcespecific models within few seconds, faster than conventional model customization approaches by orders of magnitude.
Generalization of Gutzwiller Approximation ; We derive expressions required in generalizing the Gutzwiller approximation to models comprising arbitrarily degenerate localized orbitals.
A Survey of Ultraproduct Constructions in General Topology ; We survey various attempts to transport the ultraproduct construction from the realm of model theory to that of general topology.
Generalized FroggattNielsen Mechanism ; In this paper, we propose a Generalized FroggattNielsen mechanism in which nonrenormalizable operators involving a GUT group and U1H nonsinglet Higgs field are introduced. Thus the GUT gauge symmetry breaking and the generation of hierarchical flavor hierarchy have a common origin in this mechanism. In this Generalized FroggattNielsen mechanism, we propose universality conditions for coefficients corresponding to different contractions in the group productions. We find that the predictions in Generalized FroggattNielsen mechanism for SU5 GUT is different to that of ordinary FroggattNielsen mechanism. Such Generalized FroggattNielsen mechanism can be used in GUT models when ordinary FroggattNielsen mechanism is no longer available. We study the application of Generalized FroggattNielsen mechanism in SO10 model. We find that realistic standard model mass hierarchy and mixings can be obtained both in SU5 and SO10 GUT models with such Generalized FroggattNielsen mechanism.
Interest Rates and Inflation ; A relation between interest rates and inflation is presented using a two component economic model and a simple general principle. Preliminary results indicate a remarkable similarity to classical economic theories, in particular that of Wicksell.
Perception Driven Texture Generation ; This paper investigates a novel task of generating texture images from perceptual descriptions. Previous work on texture generation focused on either synthesis from examples or generation from procedural models. Generating textures from perceptual attributes have not been well studied yet. Meanwhile, perceptual attributes, such as directionality, regularity and roughness are important factors for human observers to describe a texture. In this paper, we propose a joint deep network model that combines adversarial training and perceptual feature regression for texture generation, while only random noise and userdefined perceptual attributes are required as input. In this model, a preliminary trained convolutional neural network is essentially integrated with the adversarial framework, which can drive the generated textures to possess given perceptual attributes. An important aspect of the proposed model is that, if we change one of the input perceptual features, the corresponding appearance of the generated textures will also be changed. We design several experiments to validate the effectiveness of the proposed method. The results show that the proposed method can produce high quality texture images with desired perceptual properties.
Assessing the Stylistic Properties of Neurally Generated Text in Authorship Attribution ; Recent applications of neural language models have led to an increased interest in the automatic generation of natural language. However impressive, the evaluation of neurally generated text has so far remained rather informal and anecdotal. Here, we present an attempt at the systematic assessment of one aspect of the quality of neurally generated text. We focus on a specific aspect of neural language generation its ability to reproduce authorial writing styles. Using established models for authorship attribution, we empirically assess the stylistic qualities of neurally generated text. In comparison to conventional language models, neural models generate fuzzier text that is relatively harder to attribute correctly. Nevertheless, our results also suggest that neurally generated text offers more valuable perspectives for the augmentation of training data.
Connection formulas for the lambda generalized Ising correlation functions ; We derive and prove the connection formulas for the lambda generalized diagonal Ising model correlation functions.
Actions Generation from Captions ; Sequence transduction models have been widely explored in many natural language processing tasks. However, the target sequence usually consists of discrete tokens which represent word indices in a given vocabulary. We barely see the case where target sequence is composed of continuous vectors, where each vector is an element of a time series taken successively in a temporal domain. In this work, we introduce a new data set, named Action Generation Data Set AGDS which is specifically designed to carry out the task of captiontoaction generation. This data set contains captionaction pairs. The caption is comprised of a sequence of words describing the interactive movement between two people, and the action is a captured sequence of poses representing the movement. This data set is introduced to study the ability of generating continuous sequences through sequence transduction models. We also propose a model to innovatively combine MultiHead Attention MHA and Generative Adversarial Network GAN together. In our model, we have one generator to generate actions from captions and three discriminators where each of them is designed to carry out a unique functionality captionaction consistency discriminator, pose discriminator and pose transition discriminator. This novel design allowed us to achieve plausible generation performance which is demonstrated in the experiments.
The Generalized Crofoot Transform ; We introduce a generalized Crofoot transform between the model spaces corresponding to matrixvalued inner functions. As an application, we obtain results about matrixvalued truncated Toeplitz operators.
CookGAN Meal Image Synthesis from Ingredients ; In this work we propose a new computational framework, based on generative deep models, for synthesis of photorealistic food meal images from textual list of its ingredients. Previous works on synthesis of images from text typically rely on pretrained text models to extract text features, followed by generative neural networks GAN aimed to generate realistic images conditioned on the text features. These works mainly focus on generating spatially compact and welldefined categories of objects, such as birds or flowers, but meal images are significantly more complex, consisting of multiple ingredients whose appearance and spatial qualities are further modified by cooking methods. To generate reallike meal images from ingredients, we propose Cook Generative Adversarial Networks CookGAN, CookGAN first builds an attentionbased ingredientsimage association model, which is then used to condition a generative neural network tasked with synthesizing meal images. Furthermore, a cycleconsistent constraint is added to further improve image quality and control appearance. Experiments show our model is able to generate meal images corresponding to the ingredients.
Generating Single Peaked Votes ; We discuss how to generate singled peaked votes uniformly from the Impartial Culture model.
Generative Models for Simulating Mobility Trajectories ; Mobility datasets are fundamental for evaluating algorithms pertaining to geographic information systems and facilitating experimental reproducibility. But privacy implications restrict sharing such datasets, as even aggregated locationdata is vulnerable to membership inference attacks. Current synthetic mobility dataset generators attempt to superficially match a priori modeled mobility characteristics which do not accurately reflect the realworld characteristics. Modeling human mobility to generate synthetic yet semantically and statistically realistic trajectories is therefore crucial for publishing trajectory datasets having satisfactory utility level while preserving user privacy. Specifically, longrange dependencies inherent to human mobility are challenging to capture with both discriminative and generative models. In this paper, we benchmark the performance of recurrent neural architectures RNNs, generative adversarial networks GANs and nonparametric copulas to generate synthetic mobility traces. We evaluate the generated trajectories with respect to their geographic and semantic similarity, circadian rhythms, longrange dependencies, training and generation time. We also include two sample tests to assess statistical similarity between the observed and simulated distributions, and we analyze the privacy tradeoffs with respect to membership inference and locationsequence attacks.
Towards Understanding Generics in Mainstream OOP ; This article reports on steps towards building a simple and accurate domaintheoretic model of generic nominallytyped OOP.
Video Generation From Text ; Generating videos from text has proven to be a significant challenge for existing generative models. We tackle this problem by training a conditional generative model to extract both static and dynamic information from text. This is manifested in a hybrid framework, employing a Variational Autoencoder VAE and a Generative Adversarial Network GAN. The static features, called gist, are used to sketch textconditioned background color and object layout structure. Dynamic features are considered by transforming input text into an image filter. To obtain a large amount of data for training the deeplearning model, we develop a method to automatically create a matched textvideo corpus from publicly available online videos. Experimental results show that the proposed framework generates plausible and diverse videos, while accurately reflecting the input text information. It significantly outperforms baseline models that directly adapt texttoimage generation procedures to produce videos. Performance is evaluated both visually and by adapting the inception score used to evaluate image generation in GANs.
DANCin SEQ2SEQ Fooling Text Classifiers with Adversarial Text Example Generation ; Machine learning models are powerful but fallible. Generating adversarial examples inputs deliberately crafted to cause model misclassification or other errors can yield important insight into model assumptions and vulnerabilities. Despite significant recent work on adversarial example generation targeting image classifiers, relatively little work exists exploring adversarial example generation for text classifiers; additionally, many existing adversarial example generation algorithms require full access to target model parameters, rendering them impractical for many realworld attacks. In this work, we introduce DANCin SEQ2SEQ, a GANinspired algorithm for adversarial text example generation targeting largely blackbox text classifiers. We recast adversarial text example generation as a reinforcement learning problem, and demonstrate that our algorithm offers preliminary but promising steps towards generating semantically meaningful adversarial text examples in a realworld attack scenario.
Ambient Hidden Space of Generative Adversarial Networks ; Generative adversarial models are powerful tools to model structure in complex distributions for a variety of tasks. Current techniques for learning generative models require an access to samples which have high quality, and advanced generative models are applied to generate samples from noisy training data through ambient modules. However, the modules are only practical for the output space of the generator, and their application in the hidden space is not well studied. In this paper, we extend the ambient module to the hidden space of the generator, and provide the uniqueness condition and the corresponding strategy for the ambient hidden generator in the adversarial training process. We report the practicality of the proposed method on the benchmark dataset.
The Generalized Power Generalized Weibull Distribution Properties and Applications ; This paper introduces a new generalization of the power generalized Weibull distribution called the generalized power generalized Weibull distribution. This distribution can also be considered as a generalization of Weibull distribution. The hazard rate function of the new model has nice and flexible properties and it can take various shapes, including increasing, decreasing, upsidedown bathtub and bathtub shapes. Some of the statistical properties of the new model, including quantile function, moment generating function, reliability function, hazard function and the reverse hazard function are obtained. The moments, incomplete moments, mean deviations and Bonferroni and Lorenz curves and the order statistics densities are also derived. The model parameters are estimated by the maximum likelihood method. The usefulness of the proposed model is illustrated by using two applications of reallife data.
GPTbased Generation for Classical Chinese Poetry ; We present a simple yet effective method for generating high quality classical Chinese poetry with Generative Pretrained Language Model GPT. The method adopts a simple GPT model, without using any human crafted rules or features, or designing any additional neural components. While the proposed model learns to generate various forms of classical Chinese poems, including Jueju, Lushi, various Cipai and Couples, the generated poems are of very high quality. We also propose and implement a method to finetune the model to generate acrostic poetry. To the best of our knowledge, this is the first to employ GPT in developing a poetry generation system. We have released an online mini demonstration program on Wechat to show the generation capability of the proposed method for classical Chinese poetry.
GraphEBM Molecular Graph Generation with EnergyBased Models ; We note that most existing approaches for molecular graph generation fail to guarantee the intrinsic property of permutation invariance, resulting in unexpected bias in generative models. In this work, we propose GraphEBM to generate molecular graphs using energybased models. In particular, we parameterize the energy function in a permutation invariant manner, thus making GraphEBM permutation invariant. We apply Langevin dynamics to train the energy function by approximately maximizing likelihood and generate samples with low energies. Furthermore, to generate molecules with a desirable property, we propose a simple yet effective strategy, which pushes down energies with flexible degrees according to the properties of corresponding molecules. Finally, we explore the use of GraphEBM for generating molecules with multiple objectives in a compositional manner. Comprehensive experimental results on random, goaldirected, and compositional generation tasks demonstrate the effectiveness of our proposed method.
Generating Text through Adversarial Training using SkipThought Vectors ; GANs have been shown to perform exceedingly well on tasks pertaining to image generation and style transfer. In the field of language modelling, word embeddings such as GLoVe and word2vec are stateoftheart methods for applying neural network models on textual data. Attempts have been made to utilize GANs with word embeddings for text generation. This study presents an approach to text generation using SkipThought sentence embeddings with GANs based on gradient penalty functions and fmeasures. The proposed architecture aims to reproduce writing style in the generated text by modelling the way of expression at a sentence level across all the works of an author. Extensive experiments were run in different embedding settings on a variety of tasks including conditional text generation and language generation. The model outperforms baseline text generation networks across several automated evaluation metrics like BLEUn, METEOR and ROUGE. Further, wide applicability and effectiveness in real life tasks are demonstrated through human judgement scores.
A Framework for Automated Popsong Melody Generation with Piano Accompaniment Arrangement ; We contribute a popsong automation framework for lead melody generation and accompaniment arrangement. The framework reflects the major procedures of human music composition, generating both lead melody and piano accompaniment by a unified strategy. Specifically, we take chord progression as an input and propose three models to generate a structured melody with piano accompaniment textures. First, the harmony alternation model transforms a raw input chord progression to an altered one to better fit the specified music style. Second, the melody generation model generates the lead melody and other voices melody lines of the accompaniment using seasonal ARMA Autoregressive Moving Average processes. Third, the melody integration model integrates melody lines voices together as the final piano accompaniment. We evaluate the proposed framework using subjective listening tests. Experimental results show that the generated melodies are rated significantly higher than the ones generated by bidirectional LSTM, and our accompaniment arrangement result is comparable with a stateoftheart commercial software, Band in a Box.
Semantic Preserving Generative Adversarial Models ; We introduce generative adversarial models in which the discriminator is replaced by a calibrated nondifferentiable classifier repeatedly enhanced by domain relevant features. The role of the classifier is to prove that the actual and generated data differ over a controlled semantic space. We demonstrate that such models have the ability to generate objects with strong guarantees on their properties in a wide range of domains. They require less data than ordinary GANs, provide natural stopping conditions, uncover important properties of the data, and enhance transfer learning. Our techniques can be combined with standard generative models. We demonstrate the usefulness of our approach by applying it to several unrelated domains generating good locations for cellular antennae, molecule generation preserving key chemical properties, and generating and extrapolating lines from very few data points. Intriguing open problems are presented as well.
Generating a Common Question from Multiple Documents using Multisource EncoderDecoder Models ; Ambiguous user queries in search engines result in the retrieval of documents that often span multiple topics. One potential solution is for the search engine to generate multiple refined queries, each of which relates to a subset of the documents spanning the same topic. A preliminary step towards this goal is to generate a question that captures common concepts of multiple documents. We propose a new task of generating common question from multiple documents and present simple variant of an existing multisource encoderdecoder framework, called the MultiSource Question Generator MSQG. We first train an RNNbased single encoderdecoder generator from single document, question pairs. At test time, given multiple documents, the 'Distribute' step of our MSQG model predicts target word distributions for each document using the trained model. The 'Aggregate' step aggregates these distributions to generate a common question. This simple yet effective strategy significantly outperforms several existing baseline models applied to the new task when evaluated using automated metrics and human judgments on the MSMARCOQA dataset.
Data Augmentation for Skin Lesion using SelfAttention based Progressive Generative Adversarial Network ; Deep Neural Networks DNNs show a significant impact on medical imaging. One significant problem with adopting DNNs for skin cancer classification is that the class frequencies in the existing datasets are imbalanced. This problem hinders the training of robust and wellgeneralizing models. Data Augmentation addresses this by using existing data more effectively. However, standard data augmentation implementations are manually designed and produce only limited reasonably alternative data. Instead, Generative Adversarial Networks GANs is utilized to generate a much broader set of augmentations. This paper proposes a novel enhancement for the progressive generative adversarial networks PGAN using selfattention mechanism. Selfattention mechanism is used to directly model the longrange dependencies in the feature maps. Accordingly, selfattention complements PGAN to generate finegrained samples that comprise clinicallymeaningful information. Moreover, the stabilization technique was applied to the enhanced generative model. To train the generative models, ISIC 2018 skin lesion challenge dataset was used to synthesize highly realistic skin lesion samples for boosting further the classification result. We achieve an accuracy of 70.1 which is 2.8 better than the nonaugmented one of 67.3.
POINTER Constrained Progressive Text Generation via Insertionbased Generative Pretraining ; Largescale pretrained language models, such as BERT and GPT2, have achieved excellent performance in language representation learning and freeform text generation. However, these models cannot be directly employed to generate text under specified lexical constraints. To address this challenge, we present POINTER PrOgressive INsertionbased TransformER, a simple yet novel insertionbased approach for hardconstrained text generation. The proposed method operates by progressively inserting new tokens between existing tokens in a parallel manner. This procedure is recursively applied until a sequence is completed. The resulting coarsetofine hierarchy makes the generation process intuitive and interpretable. We pretrain our model with the proposed progressive insertionbased objective on a 12GB Wikipedia dataset, and finetune it on downstream hardconstrained generation tasks. Nonautoregressive decoding yields an empirically logarithmic time complexity during inference time. Experimental results on both News and Yelp datasets demonstrate that POINTER achieves stateoftheart performance on constrained text generation. We released the pretrained models and the source code to facilitate future research httpsgithub.comdreasysnailPOINTER.
QuasiPeriodic Parallel WaveGAN Vocoder A Nonautoregressive Pitchdependent Dilated Convolution Model for Parametric Speech Generation ; In this paper, we propose a parallel WaveGAN PWGlike neural vocoder with a quasiperiodic QP architecture to improve the pitch controllability of PWG. PWG is a compact nonautoregressive nonAR speech generation model, whose generative speed is much faster than real time. While utilizing PWG as a vocoder to generate speech on the basis of acoustic features such as spectral and prosodic features, PWG generates highfidelity speech. However, when the input acoustic features include unseen pitches, the pitch accuracy of PWGgenerated speech degrades because of the fixed and generic network of PWG without prior knowledge of speech periodicity. The proposed QPPWG adopts a pitchdependent dilated convolution network PDCNN module, which introduces the pitch information into PWG via the dynamically changed network architecture, to improve the pitch controllability and speech modeling capability of vanilla PWG. Both objective and subjective evaluation results show the higher pitch accuracy and comparable speech quality of QPPWGgenerated speech when the QPPWG model size is only 70 of that of vanilla PWG.
Network Bending Expressive Manipulation of Deep Generative Models ; We introduce a new framework for manipulating and interacting with deep generative models that we call network bending. We present a comprehensive set of deterministic transformations that can be inserted as distinct layers into the computational graph of a trained generative neural network and applied during inference. In addition, we present a novel algorithm for analysing the deep generative model and clustering features based on their spatial activation maps. This allows features to be grouped together based on spatial similarity in an unsupervised fashion. This results in the meaningful manipulation of sets of features that correspond to the generation of a broad array of semantically significant features of the generated images. We outline this framework, demonstrating our results on stateoftheart deep generative models trained on several image datasets. We show how it allows for the direct manipulation of semantically meaningful aspects of the generative process as well as allowing for a broad range of expressive outcomes.
In Search of Lost Domain Generalization ; The goal of domain generalization algorithms is to predict well on distributions different from those seen during training. While a myriad of domain generalization algorithms exist, inconsistencies in experimental conditions datasets, architectures, and model selection criteria render fair and realistic comparisons difficult. In this paper, we are interested in understanding how useful domain generalization algorithms are in realistic settings. As a first step, we realize that model selection is nontrivial for domain generalization tasks. Contrary to prior work, we argue that domain generalization algorithms without a model selection strategy should be regarded as incomplete. Next, we implement DomainBed, a testbed for domain generalization including seven multidomain datasets, nine baseline algorithms, and three model selection criteria. We conduct extensive experiments using DomainBed and find that, when carefully implemented, empirical risk minimization shows stateoftheart performance across all datasets. Looking forward, we hope that the release of DomainBed, along with contributions from fellow researchers, will streamline reproducible and rigorous research in domain generalization.
Learning to Generate Novel Domains for Domain Generalization ; This paper focuses on domain generalization DG, the task of learning from multiple source domains a model that generalizes well to unseen domains. A main challenge for DG is that the available source domains often exhibit limited diversity, hampering the model's ability to learn to generalize. We therefore employ a data generator to synthesize data from pseudonovel domains to augment the source domains. This explicitly increases the diversity of available training domains and leads to a more generalizable model. To train the generator, we model the distribution divergence between source and synthesized pseudonovel domains using optimal transport, and maximize the divergence. To ensure that semantics are preserved in the synthesized data, we further impose cycleconsistency and classification losses on the generator. Our method, L2AOT Learning to Augment by Optimal Transport outperforms current stateoftheart DG methods on four benchmark datasets.
CaMGenCausallyaware Metricguided Text Generation ; Content is created for a welldefined purpose, often described by a metric or signal represented in the form of structured information. The relationship between the goal metrics of target content and the content itself is nontrivial. While largescale language models show promising text generation capabilities, guiding the generated text with external metrics is challenging. These metrics and content tend to have inherent relationships and not all of them may be of consequence. We introduce CaMGen Causally aware Generative Networks guided by userdefined target metrics incorporating the causal relationships between the metric and content features. We leverage causal inference techniques to identify causally significant aspects of a text that lead to the target metric and then explicitly guide generative models towards these by a feedback mechanism. We propose this mechanism for variational autoencoder and Transformerbased generative models. The proposed models beat baselines in terms of the target metric control while maintaining fluency and language quality of the generated text. To the best of our knowledge, this is one of the early attempts at controlled generation incorporating a metric guide using causal inference.
Multilingual AMRtoText Generation ; Generating text from structured data is challenging because it requires bridging the gap between i structure and natural language NL and ii semantically underspecified input and fully specified NL output. Multilingual generation brings in an additional challenge that of generating into languages with varied word order and morphological properties. In this work, we focus on Abstract Meaning Representations AMRs as structured input, where previous research has overwhelmingly focused on generating only into English. We leverage advances in crosslingual embeddings, pretraining, and multilingual models to create multilingual AMRtotext models that generate in twenty one different languages. For eighteen languages, based on automatic metrics, our multilingual models surpass baselines that generate into a single language. We analyse the ability of our multilingual models to accurately capture morphology and word order using human evaluation, and find that native speakers judge our generations to be fluent.
Improving the Fairness of Deep Generative Models without Retraining ; Generative Adversarial Networks GANs advance face synthesis through learning the underlying distribution of observed data. Despite the highquality generated faces, some minority groups can be rarely generated from the trained models due to a biased image generation process. To study the issue, we first conduct an empirical study on a pretrained face synthesis model. We observe that after training the GAN model not only carries the biases in the training data but also amplifies them to some degree in the image generation process. To further improve the fairness of image generation, we propose an interpretable baseline method to balance the output facial attributes without retraining. The proposed method shifts the interpretable semantic distribution in the latent space for a more balanced image generation while preserving the sample diversity. Besides producing more balanced data regarding a particular attribute e.g., race, gender, etc., our method is generalizable to handle more than one attribute at a time and synthesize samples of finegrained subgroups. We further show the positive applicability of the balanced data sampled from GANs to quantify the biases in other face recognition systems, like commercial face attribute classifiers and face superresolution algorithms.
Learning to design druglike molecules in threedimensional space using deep generative models ; Recently, deep generative models for molecular graphs are gaining more and more attention in the field of de novo drug design. A variety of models have been developed to generate topological structures of druglike molecules, but explorations in generating threedimensional structures are still limited. Existing methods have either focused on low molecular weight compounds without considering druglikeness or generate 3D structures indirectly using atom density maps. In this work, we introduce Ligand Neural Network LNet, a novel graph generative model for designing druglike molecules with highquality 3D structures. LNet directly outputs the topological and 3D structure of molecules including hydrogen atoms, without the need for additional atom placement or bond order inference algorithm. The architecture of LNet is specifically optimized for druglike molecules, and a set of metrics is assembled to comprehensively evaluate its performance. The results show that LNet is capable of generating chemically correct, conformationally valid, and highly druglike molecules. Finally, to demonstrate its potential in structurebased molecular design, we combine LNet with MCTS and test its ability to generate potential inhibitors targeting ABL1 kinase.
Local Explanation of Dialogue Response Generation ; In comparison to the interpretation of classification models, the explanation of sequence generation models is also an important problem, however it has seen little attention. In this work, we study modelagnostic explanations of a representative text generation task dialogue response generation. Dialog response generation is challenging with its openended sentences and multiple acceptable responses. To gain insights into the reasoning process of a generation model, we propose a new method, local explanation of response generation LERG that regards the explanations as the mutual interaction of segments in input and output sentences. LERG views the sequence prediction as uncertainty estimation of a human response and then creates explanations by perturbing the input and calculating the certainty change over the human response. We show that LERG adheres to desired properties of explanations for text generation including unbiased approximation, consistency and cause identification. Empirically, our results show that our method consistently improves other widely used methods on proposed automatic and human evaluation metrics for this new task by 4.412.8. Our analysis demonstrates that LERG can extract both explicit and implicit relations between input and output segments.
Constrained Graphic Layout Generation via Latent Optimization ; It is common in graphic design humans visually arrange various elements according to their design intent and semantics. For example, a title text almost always appears on top of other elements in a document. In this work, we generate graphic layouts that can flexibly incorporate such design semantics, either specified implicitly or explicitly by a user. We optimize using the latent space of an offtheshelf layout generation model, allowing our approach to be complementary to and used with existing layout generation models. Our approach builds on a generative layout model based on a Transformer architecture, and formulates the layout generation as a constrained optimization problem where design constraints are used for element alignment, overlap avoidance, or any other userspecified relationship. We show in the experiments that our approach is capable of generating realistic layouts in both constrained and unconstrained generation tasks with a single model. The code is available at httpsgithub.comktrk115constlayout .
Multimodal Conditionality for Natural Language Generation ; Large scale pretrained language models have demonstrated stateoftheart performance in language understanding tasks. Their application has recently expanded into multimodality learning, leading to improved representations combining vision and language. However, progress in adapting language models towards conditional Natural Language Generation NLG has been limited to a single modality, generally text. We propose MAnTiS, Multimodal Adaptation for Text Synthesis, a general approach for multimodal conditionality in transformerbased NLG models. In this method, we pass inputs from each modality through modalityspecific encoders, project to textual token space, and finally join to form a conditionality prefix. We finetune the pretrained language model and encoders with the conditionality prefix guiding the generation. We apply MAnTiS to the task of product description generation, conditioning a network on both product images and titles to generate descriptive text. We demonstrate that MAnTiS outperforms strong baseline approaches on standard NLG scoring metrics. Furthermore, qualitative assessments demonstrate that MAnTiS can generate human quality descriptions consistent with given multimodal inputs.
KFCNet Knowledge Filtering and Contrastive Learning Network for Generative Commonsense Reasoning ; Pretrained language models have led to substantial gains over a broad range of natural language processing NLP tasks, but have been shown to have limitations for natural language generation tasks with highquality requirements on the output, such as commonsense generation and ad keyword generation. In this work, we present a novel Knowledge Filtering and Contrastive learning Network KFCNet which references external knowledge and achieves better generation performance. Specifically, we propose a BERTbased filter model to remove lowquality candidates, and apply contrastive learning separately to each of the encoder and decoder, within a general encoderdecoder architecture. The encoder contrastive module helps to capture global target semantics during encoding, and the decoder contrastive module enhances the utility of retrieved prototypes while learning general features. Extensive experiments on the CommonGen benchmark show that our model outperforms the previous state of the art by a large margin 6.6 points 42.5 vs. 35.9 for BLEU4, 3.7 points 33.3 vs. 29.6 for SPICE, and 1.3 points 18.3 vs. 17.0 for CIDEr. We further verify the effectiveness of the proposed contrastive module on ad keyword generation, and show that our model has potential commercial value.
Improving Compositional Generalization with SelfTraining for DatatoText Generation ; Datatotext generation focuses on generating fluent natural language responses from structured meaning representations MRs. Such representations are compositional and it is costly to collect responses for all possible combinations of atomic meaning schemata, thereby necessitating fewshot generalization to novel MRs. In this work, we systematically study the compositional generalization of the stateoftheart T5 models in fewshot datatotext tasks. We show that T5 models fail to generalize to unseen MRs, and we propose a templatebased input representation that considerably improves the model's generalization capability. To further improve the model's performance, we propose an approach based on selftraining using finetuned BLEURT for pseudo response selection. On the commonlyused SGD and Weather benchmarks, the proposed selftraining approach improves tree accuracy by 46 and reduces the slot error rates by 73 over the strong T5 baselines in fewshot settings.
Automatic Knowledge Augmentation for Generative Commonsense Reasoning ; Generative commonsense reasoning is the capability of a language model to generate a sentence with a given conceptset that is based on commonsense knowledge. However, generative language models still struggle to provide outputs, and the training set does not contain patterns that are sufficient for generative commonsense reasoning. In this paper, we propose a datacentric method that uses automatic knowledge augmentation to extend commonsense knowledge using a machine knowledge generator. This method can generate semigolden sentences that improve the generative commonsense reasoning of a language model without architecture modifications. Furthermore, this approach is a modelagnostic method and does not require human effort for data construction.
How to Build Robust FAQ Chatbot with Controllable Question Generator ; Many unanswerable adversarial questions fool the questionanswer QA system with some plausible answers. Building a robust, frequently asked questions FAQ chatbot needs a large amount of diverse adversarial examples. Recent question generation methods are ineffective at generating many highquality and diverse adversarial questionanswer pairs from unstructured text. We propose the diversity controllable semantically valid adversarial attacker DCSA, a highquality, diverse, controllable method to generate standard and adversarial samples with a semantic graph. The fluent and semantically generated QA pairs fool our passage retrieval model successfully. After that, we conduct a study on the robustness and generalization of the QA model with generated QA pairs among different domains. We find that the generated data set improves the generalizability of the QA model to the new target domain and the robustness of the QA model to detect unanswerable adversarial questions.
Boosting Generative ZeroShot Learning by Synthesizing Diverse Features with Attribute Augmentation ; The recent advance in deep generative models outlines a promising perspective in the realm of ZeroShot Learning ZSL. Most generative ZSL methods use category semantic attributes plus a Gaussian noise to generate visual features. After generating unseen samples, this family of approaches effectively transforms the ZSL problem into a supervised classification scheme. However, the existing models use a single semantic attribute, which contains the complete attribute information of the category. The generated data also carry the complete attribute information, but in reality, visual samples usually have limited attributes. Therefore, the generated data from attribute could have incomplete semantics. Based on this fact, we propose a novel framework to boost ZSL by synthesizing diverse features. This method uses augmented semantic attributes to train the generative model, so as to simulate the real distribution of visual features. We evaluate the proposed model on four benchmark datasets, observing significant performance improvement against the stateoftheart.
BioBART Pretraining and Evaluation of A Biomedical Generative Language Model ; Pretrained language models have served as important backbones for natural language processing. Recently, indomain pretraining has been shown to benefit various domainspecific downstream tasks. In the biomedical domain, natural language generation NLG tasks are of critical importance, while understudied. Approaching natural language understanding NLU tasks as NLG achieves satisfying performance in the general domain through constrained language generation or language prompting. We emphasize the lack of indomain generative language models and the unsystematic generative downstream benchmarks in the biomedical domain, hindering the development of the research community. In this work, we introduce the generative language model BioBART that adapts BART to the biomedical domain. We collate various biomedical language generation tasks including dialogue, summarization, entity linking, and named entity recognition. BioBART pretrained on PubMed abstracts has enhanced performance compared to BART and set strong baselines on several tasks. Furthermore, we conduct ablation studies on the pretraining tasks for BioBART and find that sentence permutation has negative effects on downstream tasks.
Uniform Complexity for Text Generation ; Large pretrained language models have shown promising results in a wide array of tasks such as narrative generation, question answering, and machine translation. Likewise, the current trend in literature has deeply focused on controlling salient properties of generated texts including sentiment, topic, and coherence to produce more humanlike outputs. In this work, we introduce Uniform Complexity for Text Generation or UCTG which serves as a challenge to make existing models generate uniformly complex text with respect to inputs or prompts used. For example, if the reading level of an input text prompt is appropriate for lowleveled learners ex. A2 in the CEFR, then the generated text by an NLG system should also assume this particular level for increased readability. In a controlled narrative generation task, we surveyed over 160 linguistic and cognitivelymotivated features for evaluating text readability and found out that GPT2 models and even humans struggle in preserving the linguistic complexity of input prompts used. Ultimately, we lay down potential methods and approaches which can be incorporated into the general framework of steering language models towards addressing this important challenge.
Design Target Achievement Index A Differentiable Metric to Enhance Deep Generative Models in MultiObjective Inverse Design ; Deep Generative Machine Learning Models have been growing in popularity across the design community thanks to their ability to learn and mimic complex data distributions. While early works are promising, further advancement will depend on addressing several critical considerations such as design quality, feasibility, novelty, and targeted inverse design. We propose the Design Target Achievement Index DTAI, a differentiable, tunable metric that scores a design's ability to achieve designerspecified minimum performance targets. We demonstrate that DTAI can drastically improve the performance of generated designs when directly used as a training loss in Deep Generative Models. We apply the DTAI loss to a PerformanceAugmented Diverse GAN PaDGAN and demonstrate superior generative performance compared to a set of baseline Deep Generative Models including a MultiObjective PaDGAN and specialized tabular generation algorithms like the Conditional Tabular GAN CTGAN. We further enhance PaDGAN with an auxiliary feasibility classifier to encourage feasible designs. To evaluate methods, we propose a comprehensive set of evaluation metrics for generative methods that focus on feasibility, diversity, and satisfaction of design performance targets. Methods are tested on a challenging benchmarking problem the FRAMED bicycle frame design dataset featuring mixeddatatype parametric data, heavily skewed and multimodal distributions, and ten competing performance objectives.
Evaluating the Generalization Ability of SuperResolution Networks ; Performance and generalization ability are two important aspects to evaluate the deep learning models. However, research on the generalization ability of SuperResolution SR networks is currently absent. Assessing the generalization ability of deep models not only helps us to understand their intrinsic mechanisms, but also allows us to quantitatively measure their applicability boundaries, which is important for unrestricted realworld applications. To this end, we make the first attempt to propose a Generalization Assessment Index for SR networks, namely SRGA. SRGA exploits the statistical characteristics of the internal features of deep networks to measure the generalization ability. Specially, it is a nonparametric and nonlearning metric. To better validate our method, we collect a patchbased image evaluation set PIES that includes both synthetic and realworld images, covering a wide range of degradations. With SRGA and PIES dataset, we benchmark existing SR models on the generalization ability. This work provides insights and tools for future research on model generalization in lowlevel vision.
RSTGen Imbuing FineGrained Interpretable Control into LongFormText Generators ; In this paper, we study the task of improving the cohesion and coherence of longform text generated by language models. To this end, we propose RSTGen, a framework that utilises Rhetorical Structure Theory RST, a classical language theory, to control the discourse structure, semantics and topics of generated text. Firstly, we demonstrate our model's ability to control structural discourse and semantic features of generated text in open generation evaluation. Then we experiment on the two challenging longform text tasks of argument generation and story generation. Evaluation using automated metrics and a metric with high correlation to human evaluation, shows that our model performs competitively against existing models, while offering significantly more controls over generated text than alternative methods.
When happy accidents spark creativity Bringing collaborative speculation to life with generative AI ; Generative AI techniques like those that synthesize images from text texttoimage models offer new possibilities for creatively imagining new ideas. We investigate the capabilities of these models to help communities engage in conversations about their collective future. In particular, we design and deploy a facilitated experience where participants collaboratively speculate on utopias they want to see, and then produce AIgenerated imagery from those speculations. In a series of indepth user interviews, we invite participants to reflect on the generated images and refine their visions for the future. We synthesize findings with a bespoke community zine on the experience. We observe that participants often generated ideas for implementing their vision and drew new lateral considerations as a result of viewing the generated images. Critically, we find that the unexpected difference between the participant's imagined output and the generated image is what facilitated new insight for the participant. We hope our experimental model for cocreation, computational creativity, and community reflection inspires the use of generative models to help communities and organizations envision better futures.
Language Models are Realistic Tabular Data Generators ; Tabular data is among the oldest and most ubiquitous forms of data. However, the generation of synthetic samples with the original data's characteristics remains a significant challenge for tabular data. While many generative models from the computer vision domain, such as variational autoencoders or generative adversarial networks, have been adapted for tabular data generation, less research has been directed towards recent transformerbased large language models LLMs, which are also generative in nature. To this end, we propose GReaT Generation of Realistic Tabular data, which exploits an autoregressive generative LLM to sample synthetic and yet highly realistic tabular data. Furthermore, GReaT can model tabular data distributions by conditioning on any subset of features; the remaining features are sampled without additional overhead. We demonstrate the effectiveness of the proposed approach in a series of experiments that quantify the validity and quality of the produced data samples from multiple angles. We find that GReaT maintains stateoftheart performance across numerous realworld and synthetic data sets with heterogeneous feature types coming in various sizes.
P3LM Probabilistically Permuted Prophet Language Modeling for Generative PreTraining ; Conventional autoregressive lefttoright L2R sequence generation faces two issues during decoding limited to unidirectional target sequence modeling, and constrained on strong local dependencies. To address the aforementioned problem, we propose P3LM, a probabilistically permuted prophet language model, which strengthens the modeling of bidirectional information and long token dependencies for sequence generation. Specifically, P3LM learns to generate tokens in permuted order upon an orderaware transformer decoder, as well as to generate the corresponding future N tokens with a multistream attention mechanism. Extensive experiments are conducted on the GLGE benchmark, which includes four datasets for summarization, two for question generation, one for conversational question answering, and one for dialog response generation, where P3LM achieves stateoftheart results compared with strong publicly available generative pretraining methods.
CELLS A Parallel Corpus for Biomedical Lay Language Generation ; Recent lay language generation systems have used Transformer models trained on a parallel corpus to increase health information accessibility. However, the applicability of these models is constrained by the limited size and topical breadth of available corpora. We introduce CELLS, the largest 63k pairs and broadestranging 12 journals parallel corpus for lay language generation. The abstract and the corresponding lay language summary are written by domain experts, assuring the quality of our dataset. Furthermore, qualitative evaluation of expertauthored plain language summaries has revealed background explanation as a key strategy to increase accessibility. Such explanation is challenging for neural models to generate because it goes beyond simplification by adding content absent from the source. We derive two specialized paired corpora from CELLS to address key challenges in lay language generation generating background explanations and simplifying the original abstract. We adopt retrievalaugmented models as an intuitive fit for the task of background explanation generation, and show improvements in summary quality and simplicity while maintaining factual correctness. Taken together, this work presents the first comprehensive study of background explanation for lay language generation, paving the path for disseminating scientific knowledge to a broader audience. CELLS is publicly available at httpsgithub.comLinguisticAnomaliesplsretrieval.
Reasoning Circuits Fewshot Multihop Question Generation with Structured Rationales ; Multihop Question Generation is the task of generating questions which require the reader to reason over and combine information spread across multiple passages using several reasoning steps. Chainofthought rationale generation has been shown to improve performance on multistep reasoning tasks and make model predictions more interpretable. However, fewshot performance gains from including rationales have been largely observed only in 100B language models, and otherwise require large scale manual rationale annotation. In this work, we introduce a new framework for applying chainofthought inspired structured rationale generation to multihop question generation under a very low supervision regime 8 to 128shot. We propose to annotate a small number of examples following our proposed multistep rationale schema, treating each reasoning step as a separate task to be performed by a generative language model. We show that our framework leads to improved control over the difficulty of the generated questions and better performance compared to baselines trained without rationales, both on automatic evaluation metrics and in human evaluation. Importantly, we show that this is achievable with a modest model size.
Effective Dynamics of Generative Adversarial Networks ; Generative adversarial networks GANs are a class of machinelearning models that use adversarial training to generate new samples with the same potentially very complex statistics as the training samples. One major form of training failure, known as mode collapse, involves the generator failing to reproduce the full diversity of modes in the target probability distribution. Here, we present an effective model of GAN training, which captures the learning dynamics by replacing the generator neural network with a collection of particles in the output space; particles are coupled by a universal kernel valid for certain wide neural networks and highdimensional inputs. The generality of our simplified model allows us to study the conditions under which mode collapse occurs. Indeed, experiments which vary the effective kernel of the generator reveal a mode collapse transition, the shape of which can be related to the type of discriminator through the frequency principle. Further, we find that gradient regularizers of intermediate strengths can optimally yield convergence through critical damping of the generator dynamics. Our effective GAN model thus provides an interpretable physical framework for understanding and improving adversarial training.
Modified Query Expansion Through Generative Adversarial Networks for Information Extraction in ECommerce ; This work addresses an alternative approach for query expansion QE using a generative adversarial network GAN to enhance the effectiveness of information search in ecommerce. We propose a modified QE conditional GAN mQECGAN framework, which resolves keywords by expanding the query with a synthetically generated query that proposes semantic information from text input. We train a sequencetosequence transformer model as the generator to produce keywords and use a recurrent neural network model as the discriminator to classify an adversarial output with the generator. With the modified CGAN framework, various forms of semantic insights gathered from the query document corpus are introduced to the generation process. We leverage these insights as conditions for the generator model and discuss their effectiveness for the query expansion task. Our experiments demonstrate that the utilization of condition structures within the mQECGAN framework can increase the semantic similarity between generated sequences and reference documents up to nearly 10 compared to baseline models
3DShape2VecSet A 3D Shape Representation for Neural Fields and Generative Diffusion Models ; We introduce 3DShape2VecSet, a novel shape representation for neural fields designed for generative diffusion models. Our shape representation can encode 3D shapes given as surface models or point clouds, and represents them as neural fields. The concept of neural fields has previously been combined with a global latent vector, a regular grid of latent vectors, or an irregular grid of latent vectors. Our new representation encodes neural fields on top of a set of vectors. We draw from multiple concepts, such as the radial basis function representation and the cross attention and selfattention function, to design a learnable representation that is especially suitable for processing with transformers. Our results show improved performance in 3D shape encoding and 3D shape generative modeling tasks. We demonstrate a wide variety of generative applications unconditioned generation, categoryconditioned generation, textconditioned generation, pointcloud completion, and imageconditioned generation.
RePrompt Automatic Prompt Editing to Refine AIGenerative Art Towards Precise Expressions ; Generative AI models have shown impressive ability to produce images with text prompts, which could benefit creativity in visual art creation and selfexpression. However, it is unclear how precisely the generated images express contexts and emotions from the input texts. We explored the emotional expressiveness of AIgenerated images and developed RePrompt, an automatic method to refine text prompts toward precise expression of the generated images. Inspired by crowdsourced editing strategies, we curated intuitive text features, such as the number and concreteness of nouns, and trained a proxy model to analyze the feature effects on the AIgenerated image. With model explanations of the proxy model, we curated a rubric to adjust text prompts to optimize image generation for precise emotion expression. We conducted simulation and user studies, which showed that RePrompt significantly improves the emotional expressiveness of AIgenerated images, especially for negative emotions.
Continuous descriptorbased control for deep audio synthesis ; Despite significant advances in deep models for music generation, the use of these techniques remains restricted to expert users. Before being democratized among musicians, generative models must first provide expressive control over the generation, as this conditions the integration of deep generative models in creative workflows. In this paper, we tackle this issue by introducing a deep generative audio model providing expressive and continuous descriptorbased control, while remaining lightweight enough to be embedded in a hardware synthesizer. We enforce the controllability of realtime generation by explicitly removing salient musical features in the latent space using an adversarial confusion criterion. Userspecified features are then reintroduced as additional conditioning information, allowing for continuous control of the generation, akin to a synthesizer knob. We assess the performance of our method on a wide variety of sounds including instrumental, percussive and speech recordings while providing both timbre and attributes transfer, allowing new ways of generating sounds.
Unifying Layout Generation with a Decoupled Diffusion Model ; Layout generation aims to synthesize realistic graphic scenes consisting of elements with different attributes including category, size, position, and betweenelement relation. It is a crucial task for reducing the burden on heavyduty graphic design works for formatted scenes, e.g., publications, documents, and user interfaces UIs. Diverse application scenarios impose a big challenge in unifying various layout generation subtasks, including conditional and unconditional generation. In this paper, we propose a Layout Diffusion Generative Model LDGM to achieve such unification with a single decoupled diffusion model. LDGM views a layout of arbitrary missing or coarse element attributes as an intermediate diffusion status from a completed layout. Since different attributes have their individual semantics and characteristics, we propose to decouple the diffusion processes for them to improve the diversity of training samples and learn the reverse process jointly to exploit globalscope contexts for facilitating generation. As a result, our LDGM can generate layouts either from scratch or conditional on arbitrary available attributes. Extensive qualitative and quantitative experiments demonstrate our proposed LDGM outperforms existing layout generation models in both functionality and performance.
GNNBuilder An Automated Framework for Generic Graph Neural Network Accelerator Generation, Simulation, and Optimization ; There are plenty of graph neural network GNN accelerators being proposed. However, they highly rely on users' hardware expertise and are usually optimized for one specific GNN model, making them challenging for practical use. Therefore, in this work, we propose GNNBuilder, the first automated, generic, endtoend GNN accelerator generation framework. It features four advantages 1 GNNBuilder can automatically generate GNN accelerators for a wide range of GNN models arbitrarily defined by users; 2 GNNBuilder takes standard PyTorch programming interface, introducing zero overhead for algorithm developers; 3 GNNBuilder supports endtoend code generation, simulation, accelerator optimization, and hardware deployment, realizing a pushbutton fashion for GNN accelerator design; 4 GNNBuilder is equipped with accurate performance models of its generated accelerator, enabling fast and flexible design space exploration DSE. In the experiments, first, we show that our accelerator performance model has errors within 36 for latency prediction and 18 for BRAM count prediction. Second, we show that our generated accelerators can outperform CPU by 6.33times and GPU by 6.87times. This framework is opensource, and the code is available at httpsgithub.comsharclabgnnbuilder.
DiffCollage Parallel Generation of Large Content with Diffusion Models ; We present DiffCollage, a compositional diffusion model that can generate large content by leveraging diffusion models trained on generating pieces of the large content. Our approach is based on a factor graph representation where each factor node represents a portion of the content and a variable node represents their overlap. This representation allows us to aggregate intermediate outputs from diffusion models defined on individual nodes to generate content of arbitrary size and shape in parallel without resorting to an autoregressive generation procedure. We apply DiffCollage to various tasks, including infinite image generation, panorama image generation, and longduration textguided motion generation. Extensive experimental results with a comparison to strong autoregressive baselines verify the effectiveness of our approach.
Token Imbalance Adaptation for Radiology Report Generation ; Imbalanced token distributions naturally exist in text documents, leading neural language models to overfit on frequent tokens. The token imbalance may dampen the robustness of radiology report generators, as complex medical terms appear less frequently but reflect more medical information. In this study, we demonstrate how current stateoftheart models fail to generate infrequent tokens on two standard benchmark datasets IU XRAY and MIMICCXR of radiology report generation. However, no prior study has proposed methods to adapt infrequent tokens for text generators feeding with medical images. To solve the challenge, we propose the textbfToken textbfImbalance Adapttextbfer textitTIMER, aiming to improve generation robustness on infrequent tokens. The model automatically leverages token imbalance by an unlikelihood loss and dynamically optimizes generation processes to augment infrequent tokens. We compare our approach with multiple stateoftheart methods on the two benchmarks. Experiments demonstrate the effectiveness of our approach in enhancing model robustness overall and infrequent tokens. Our ablation analysis shows that our reinforcement learning method has a major effect in adapting token imbalance for radiology report generation.
Directed Acyclic Transformer Pretraining for Highquality Nonautoregressive Text Generation ; NonAutoRegressive NAR text generation models have drawn much attention because of their significantly faster decoding speed and good generation quality in machine translation. However, in a wider range of text generation tasks, existing NAR models lack proper pretraining, making them still far behind the pretrained autoregressive models. In this paper, we propose Pretrained Directed Acyclic Transformer PreDAT and a novel pretraining task to promote prediction consistency in NAR generation. Experiments on five text generation tasks show that our PreDAT remarkably outperforms existing pretrained NAR models 4.2 scores on average and even achieves better results than pretrained autoregressive baselines in ngrambased metrics, along with 17 times speedup in throughput. Further analysis shows that PreDAT benefits from the unbiased prediction order that alleviates the error accumulation problem in autoregressive generation, which provides new insights into the advantages of NAR generation.
CGCE A Chinese Generative Chat Evaluation Benchmark for General and Financial Domains ; Generative chat models, such as ChatGPT and GPT4, have revolutionized natural language generation NLG by incorporating instructions and human feedback to achieve significant performance improvements. However, the lack of standardized evaluation benchmarks for chat models, particularly for Chinese and domainspecific models, hinders their assessment and progress. To address this gap, we introduce the Chinese Generative Chat Evaluation CGCE benchmark, focusing on general and financial domains. The CGCE benchmark encompasses diverse tasks, including 200 questions in the general domain and 150 specific professional questions in the financial domain. Manual scoring evaluates factors such as accuracy, coherence, expression clarity, and completeness. The CGCE benchmark provides researchers with a standardized framework to assess and compare Chinese generative chat models, fostering advancements in NLG research.
Revisit and Outstrip Entity Alignment A Perspective of Generative Models ; Recent embeddingbased methods have achieved great successes on exploiting entity alignment from knowledge graph KG embeddings of multiple modals. In this paper, we study embeddingbased entity alignment EEA from a perspective of generative models. We show that EEA is a special problem where the main objective is analogous to that in a typical generative model, based on which we theoretically prove the effectiveness of the recently developed generative adversarial network GANbased EEA methods. We then reveal that their incomplete objective limits the capacity on both entity alignment and entity synthesis i.e., generating new entities. We mitigate this problem by introducing a generative EEA abbr., GEEA framework with the proposed mutual variational autoencoder MVAE as the generative model. MVAE can convert an entity from one KG to another and generate new entities from random noise vectors. We demonstrate the power of GEEA with theoretical analysis and empirical experiments on both entity alignment and entity synthesis tasks.
Towards SymmetryAware Generation of Periodic Materials ; We consider the problem of generating periodic materials with deep models. While symmetryaware molecule generation has been studied extensively, periodic materials possess different symmetries, which have not been completely captured by existing methods. In this work, we propose SyMat, a novel material generation approach that can capture physical symmetries of periodic material structures. SyMat generates atom types and lattices of materials through generating atom type sets, lattice lengths and lattice angles with a variational autoencoder model. In addition, SyMat employs a scorebased diffusion model to generate atom coordinates of materials, in which a novel symmetryaware probabilistic model is used in the coordinate diffusion process. We show that SyMat is theoretically invariant to all symmetry transformations on materials and demonstrate that SyMat achieves promising performance on random generation and property optimization tasks.
GenAssist Making Image Generation Accessible ; Blind and low vision BLV creators use images to communicate with sighted audiences. However, creating or retrieving images is challenging for BLV creators as it is difficult to use authoring tools or assess image search results. Thus, creators limit the types of images they create or recruit sighted collaborators. While texttoimage generation models let creators generate highfidelity images based on a text description i.e. prompt, it is difficult to assess the content and quality of generated images. We present GenAssist, a system to make texttoimage generation accessible. Using our interface, creators can verify whether generated image candidates followed the prompt, access additional details in the image not specified in the prompt, and skim a summary of similarities and differences between image candidates. To power the interface, GenAssist uses a large language model to generate visual questions, visionlanguage models to extract answers, and a large language model to summarize the results. Our study with 12 BLV creators demonstrated that GenAssist enables and simplifies the process of image selection and generation, making visual authoring more accessible to all.
AdvDiff Generating Unrestricted Adversarial Examples using Diffusion Models ; Unrestricted adversarial attacks present a serious threat to deep learning models and adversarial defense techniques. They pose severe security problems for deep learning applications because they can effectively bypass defense mechanisms. However, previous attack methods often utilize Generative Adversarial Networks GANs, which are not theoretically provable and thus generate unrealistic examples by incorporating adversarial objectives, especially for largescale datasets like ImageNet. In this paper, we propose a new method, called AdvDiff, to generate unrestricted adversarial examples with diffusion models. We design two novel adversarial guidance techniques to conduct adversarial sampling in the reverse generation process of diffusion models. These two techniques are effective and stable to generate highquality, realistic adversarial examples by integrating gradients of the target classifier interpretably. Experimental results on MNIST and ImageNet datasets demonstrate that AdvDiff is effective to generate unrestricted adversarial examples, which outperforms GANbased methods in terms of attack performance and generation quality.
Detoxify Language Model StepbyStep ; Detoxification for LLMs is challenging since it requires models to avoid generating harmful content while maintaining the generation capability. To ensure the safety of generations, previous detoxification methods detoxify the models by changing the data distributions or constraining the generations from different aspects in a singlestep manner. However, these approaches will dramatically affect the generation quality of LLMs, e.g., discourse coherence and semantic consistency, since language models tend to generate along the toxic prompt while detoxification methods work in the opposite direction. To handle such a conflict, we decompose the detoxification process into different substeps, where the detoxification is concentrated in the input stage and the subsequent continual generation is based on the nontoxic prompt. Besides, we also calibrate the strong reasoning ability of LLMs by designing a DetoxChain to connect the above substeps in an orderly manner, which allows LLMs to detoxify the text stepbystep. Automatic and human evaluation on two benchmarks reveals that by training with DetoxChain, six LLMs scaling from 1B to 33B can obtain significant detoxification and generation improvement. Our code and data are available at httpsgithub.comCODINNLGDetoxCoT. Warning examples in the paper may contain uncensored offensive content.
Continual ZeroShot Learning through Semantically Guided Generative Random Walks ; Learning novel concepts, remembering previous knowledge, and adapting it to future tasks occur simultaneously throughout a human's lifetime. To model such comprehensive abilities, continual zeroshot learning CZSL has recently been introduced. However, most existing methods overused unseen semantic information that may not be continually accessible in realistic settings. In this paper, we address the challenge of continual zeroshot learning where unseen information is not provided during training, by leveraging generative modeling. The heart of the generativebased methods is to learn quality representations from seen classes to improve the generative understanding of the unseen visual space. Motivated by this, we introduce generalizationbound tools and provide the first theoretical explanation for the benefits of generative modeling to CZSL tasks. Guided by the theoretical analysis, we then propose our learning algorithm that employs a novel semantically guided Generative Random Walk GRW loss. The GRW loss augments the training by continually encouraging the model to generate realistic and characterized samples to represent the unseen space. Our algorithm achieves stateoftheart performance on AWA1, AWA2, CUB, and SUN datasets, surpassing existing CZSL methods by 37. The code has been made available here urlhttpsgithub.comwxzhangIGCZSL
DiffRetinex Rethinking Lowlight Image Enhancement with A Generative Diffusion Model ; In this paper, we rethink the lowlight image enhancement task and propose a physically explainable and generative diffusion model for lowlight image enhancement, termed as DiffRetinex. We aim to integrate the advantages of the physical model and the generative network. Furthermore, we hope to supplement and even deduce the information missing in the lowlight image through the generative network. Therefore, DiffRetinex formulates the lowlight image enhancement problem into Retinex decomposition and conditional image generation. In the Retinex decomposition, we integrate the superiority of attention in Transformer and meticulously design a Retinex Transformer decomposition network TDN to decompose the image into illumination and reflectance maps. Then, we design multipath generative diffusion networks to reconstruct the normallight Retinex probability distribution and solve the various degradations in these components respectively, including dark illumination, noise, color deviation, loss of scene contents, etc. Owing to generative diffusion model, DiffRetinex puts the restoration of lowlight subtle detail into practice. Extensive experiments conducted on realworld lowlight datasets qualitatively and quantitatively demonstrate the effectiveness, superiority, and generalization of the proposed method.
SyncDreamer Generating Multiviewconsistent Images from a Singleview Image ; In this paper, we present a novel diffusion model called that generates multiviewconsistent images from a singleview image. Using pretrained largescale 2D diffusion models, recent work Zero123 demonstrates the ability to generate plausible novel views from a singleview image of an object. However, maintaining consistency in geometry and colors for the generated images remains a challenge. To address this issue, we propose a synchronized multiview diffusion model that models the joint probability distribution of multiview images, enabling the generation of multiviewconsistent images in a single reverse process. SyncDreamer synchronizes the intermediate states of all the generated images at every step of the reverse process through a 3Daware feature attention mechanism that correlates the corresponding features across different views. Experiments show that SyncDreamer generates images with high consistency across different views, thus making it wellsuited for various 3D generation tasks such as novelviewsynthesis, textto3D, and imageto3D.
Compositional Sculpting of Iterative Generative Processes ; High training costs of generative models and the need to finetune them for specific tasks have created a strong interest in model reuse and composition. A key challenge in composing iterative generative processes, such as GFlowNets and diffusion models, is that to realize the desired target distribution, all steps of the generative process need to be coordinated, and satisfy delicate balance conditions. In this work, we propose Compositional Sculpting a general approach for defining compositions of iterative generative processes. We then introduce a method for sampling from these compositions built on classifier guidance. We showcase ways to accomplish compositional sculpting in both GFlowNets and diffusion models. We highlight two binary operations unicodex2014 the harmonic mean p1 otimes p2 and the contrast p1 unicodex25D1,p2 between pairs, and the generalization of these operations to multiple component distributions. We offer empirical results on image and molecular generation tasks.
MiniGPT5 Interleaved VisionandLanguage Generation via Generative Vokens ; Large Language Models LLMs have garnered significant attention for their advancements in natural language processing, demonstrating unparalleled prowess in text comprehension and generation. Yet, the simultaneous generation of images with coherent textual narratives remains an evolving frontier. In response, we introduce an innovative interleaved visionandlanguage generation technique anchored by the concept of generative vokens, acting as the bridge for harmonized imagetext outputs. Our approach is characterized by a distinctive twostaged training strategy focusing on descriptionfree multimodal generation, where the training requires no comprehensive descriptions of images. To bolster model integrity, classifierfree guidance is incorporated, enhancing the effectiveness of vokens on image generation. Our model, MiniGPT5, exhibits substantial improvement over the baseline Divter model on the MMDialog dataset and consistently delivers superior or comparable multimodal outputs in human evaluations on the VIST dataset, highlighting its efficacy across diverse benchmarks.
Macroscopic traffic models from microscopic carfollowing models ; We present a method to derive macroscopic fluiddynamic models from microscopic carfollowing models via a coarsegraining procedure. The method is first demonstrated for the optimal velocity model. The derived macroscopic model consists of a conservation equation and a momentum equation, and the latter contains a relaxation term, an anticipation term, and a diffusion term. Properties of the resulting macroscopic model are compared with those of the optimal velocity model through numerical simulations, and reasonable agreement is found although there are deviations in the quantitative level. The derivation is also extended to general carfollowing models.
Concurrent Models for Function Execution ; We derive an abstract computational model from a sequential computational model that is generally used for function execution. This abstract computational model allows for the concurrent execution of functions. We discuss concurrent models for function execution as implementations from the abstract computational model. We give an example of a particular concurrent function construct that can be implemented on a concurrent machine model using multithreading. The result is a framework of computational models at different levels of abstraction that can be used in further development of concurrent computational models that deal with the problems inherent with concurrency.
Generalizing and Hybridizing Countbased and Neural Language Models ; Language models LMs are statistical models that calculate probabilities over sequences of words or other discrete symbols. Currently two major paradigms for language modeling exist countbased ngram models, which have advantages of scalability and testtime speed, and neural LMs, which often achieve superior modeling performance. We demonstrate how both varieties of models can be unified in a single modeling framework that defines a set of probability distributions over the vocabulary of words, and then dynamically calculates mixture weights over these distributions. This formulation allows us to create novel hybrid models that combine the desirable features of countbased and neural LMs, and experiments demonstrate the advantages of these approaches.
What is an Ordinal Latent Trait Model ; Although various polytomous item response models are considered to be ordinal models there seems no general definition of an ordinal model available. Alternative concepts of ordinal models are discussed and it is shown that they coincide for classical unidimensional models. For multidimensional models the definition of an ordinal model refers to specific traits in the multidimensional space of traits. The objective is to provide a theoretical framework for ordinal models. Practical considerations concerning the strength of the link between the latent trait and the order of categories are considered briefly.
Learning to Discover, Ground and Use Words with Segmental Neural Language Models ; We propose a segmental neural language model that combines the generalization power of neural networks with the ability to discover wordlike units that are latent in unsegmented character sequences. In contrast to previous segmentation models that treat word segmentation as an isolated task, our model unifies word discovery, learning how words fit together to form sentences, and, by conditioning the model on visual context, how words' meanings ground in representations of nonlinguistic modalities. Experiments show that the unconditional model learns predictive distributions better than character LSTM models, discovers words competitively with nonparametric Bayesian word segmentation models, and that modeling language conditional on visual context improves performance on both.
Accurate Semidefinite Programming Models for Optimal Power Flow in Distribution Systems ; In this paper, we develop semidefinite programming SDP models aimed at solving optimal power flow OPF problems in distribution systems. We propose two models the symmetrical SDP model which modifies the existing BFMSDP model. Then based on the symmetrical SDP model, we develop a voltage regulation model that solves OPF problems with binding voltage constraints. To evaluate the accuracy of our proposed OPF models, we rely on OpenDSS, a power flow solver, to generate power flow solutions as the benchmarks. Comprehensive case studies are conducted showing our SDP models have better numerical stability and yield more accurate results than existing approaches
Transition Models for Count Data a Flexible Alternative to Fixed Distribution Models ; A flexible semiparametric class of models is introduced that offers an alternative to classical regression models for count data as the Poisson and negative binomial model, as well as to more general models accounting for excess zeros that are also based on fixed distributional assumptions. The model allows that the data itself determine the distribution of the response variable, but, in its basic form, uses a parametric term that specifies the effect of explanatory variables. In addition, an extended version is considered, in which the effects of covariates are specified nonparametrically. The proposed model and traditional models are compared by utilizing several real data applications.
Model Pluralism ; This paper introduces and defends an account of modelbased science that I dub model pluralism. I argue that despite a growing awareness in the philosophy of science literature of the multiplicity, diversity, and richness of models and modelingpractices, more radical conclusions follow from this recognition than have previously been inferred. Going against the tendency within the literature to generalize from single models, I explicate and defend the following two core theses i any successful analysis of models must target sets of models, their multiplicity of functions within science, and their scientific context and history and ii for almost any aspect x of phenomenon y, scientists require multiple models to achieve scientific goal z.
Overestimation of Syntactic Representationin Neural Language Models ; With the advent of powerful neural language models over the last few years, research attention has increasingly focused on what aspects of language they represent that make them so successful. Several testing methodologies have been developed to probe models' syntactic representations. One popular method for determining a model's ability to induce syntactic structure trains a model on strings generated according to a template then tests the model's ability to distinguish such strings from superficially similar ones with different syntax. We illustrate a fundamental problem with this approach by reproducing positive results from a recent paper with two nonsyntactic baseline language models an ngram model and an LSTM model trained on scrambled inputs.
Learning Convex Optimization Models ; A convex optimization model predicts an output from an input by solving a convex optimization problem. The class of convex optimization models is large, and includes as special cases many wellknown models like linear and logistic regression. We propose a heuristic for learning the parameters in a convex optimization model given a dataset of inputoutput pairs, using recently developed methods for differentiating the solution of a convex optimization problem with respect to its parameters. We describe three general classes of convex optimization models, maximum a posteriori MAP models, utility maximization models, and agent models, and present a numerical experiment for each.
Can Model Compression Improve NLP Fairness ; Model compression techniques are receiving increasing attention; however, the effect of compression on model fairness is still under explored. This is the first paper to examine the effect of distillation and pruning on the toxicity and bias of generative language models. We test Knowledge Distillation and Pruning methods on the GPT2 model and found a consistent pattern of toxicity and bias reduction after model distillation; this result can be potentially interpreted by existing line of research which describes model compression as a regularization technique; our work not only serves as a reference for safe deployment of compressed models, but also extends the discussion of compression as regularization into the setting of neural LMs, and hints at the possibility of using compression to develop fairer models.